The purpose of this study was to determine the in-vivo andin-vitro effects of insulin, at physiological and supraphysiological concentrations, on the human immune system. Ten healthy young men went through a sequential two-step hyperinsulinaemic euglycaemic clamp. Plasma insulin concentrations were increased from baseline (9.0 µU/ml) to 49.1µU/ml after 1 h of insulin infusion (step I) and to 1281 µU/ml(step II) after 2 h of infusion. As control experiments infusions of isotonic saline were performed. The unstimulated natural killer (NK) cell activity among blood mononuclear cells (BMNC) increased in response to supraphysiological plasma insulin levels (baseline versus step II: 20.6± 11.3 versus 27.8 ± 14.4%). The percentages of the CD16+ NK cells did not change, indicating an enhanced cytotoxic capability per individual NK cell. Insulin also slightly increased the activity of NK cells in vitro. A decline at step II in the concentrations of monocytes (0.29 ± 0.09 versus 0.12 ± 0.03× 109/L), lymphocytes (1.57 ± 0.46 versus 1.22 ± 0.25 × 109/L), and CD16+ (24.2 ± 17.5 versus 16.7 ± 11.2 × 107/L), CD14+ (20.9 ± 10.8 versus 8.6 ± 3.9 × 107/L), HLA-DR+ (37.2± 22.1 versus 19.2 ± 10.7 × 107/L) and CD45RO+ (91.6 ± 33.4 versus 61.7 ± 6.4 × 107/L) cells as well as in the percentages of CD14+ cells(11.2 ± 4.7 versus 6.4 ± 2.3%) and CD14+/HLA-DR+ monocytes (9.7 ± 3.9 versus 4.8 ± 2.8%) were observed. No changes were found at step I. Hyperinsulinaemia did not change the percentages of the CD3+, CD4+, CD8+, CD19+, CD56+, CD11a+, CD45RO+ and CD45RA+ cells, the numbers of circulating immunoglobulin (Ig)G-, IgA- and IgM- secreting cells, or the proliferative responses of BMNC to phytohaemagglutinin, purified derivative of tuberculin or interleukin (IL)-2. Hyperinsulinaemia did not change thein-vitro sensibility to insulin. In conclusion, supraphysiological insulin levels increased the activity of the individual NK cells, but decreased the numbers of NK cells, lymphocytes and activated monocytes. The findings are presumably of minor clinical relevance but may indicate an insulin-induced immune activation.